Weekly Threat Landscape Digest – Week 47

- Apache Roller XML-RPC Deserialization Remote Code Execution CVE-2026-129448 (CVSS 9.8, Actively Exploited)
Overview
- Critical unauthenticated deserialization flaw in Apache Roller’s XML-RPC endpoint lets remote attackers execute arbitrary code by submitting malicious serialized data before authentication.
- Active exploitation confirmed in the wild. CVE-2026-129448 is the same issue as CVE-2026-82384, previously noted against Apache Roller 6.1.5.
Impact
- Unauthenticated remote code execution leading to system compromise, data theft, and service disruption.
Affected / Fixed Versions
- Affected: Apache Roller 4.0.0 through 6.1.5. Fixed: 6.1.6 and later.
Recommendations
- Immediately upgrade to Apache Roller 6.1.6 or later.
- Review server logs for anomalous XML-RPC requests; monitor for unusual processes, unauthorized modifications, and suspicious network activity.
Reference Links
- F5 BIG-IP and NGINX Critical Vulnerabilities Including Arbitrary Code Execution in NGINX HTTP/2 Module CVE-2026-123007 (PoC Published)
Overview
- Multiple critical and high-severity flaws across BIG-IP (APM, DNS, PEM, SSL Orchestrator), BIG-IQ, BIG-IP Cloud-Native Edition, BIG-IP Next for Kubernetes, NGINX Open Source, NGINX Plus, NGINX Instance Manager, and related WAF and DoS modules.
- CVE-2026-123007 in the NGINX HTTP/2 module has a public proof of concept. Flaws allow code execution, DoS, memory corruption, security bypass, and system compromise.
Impact
- Remote code execution, service disruption, and full compromise of affected platforms.
Affected / Fixed Versions
- Detailed affected and fixed versions are in the official F5 advisory.
Recommendations
- Apply official F5 security updates and hotfixes; prioritize critical flaws on internet-facing systems.
- Audit configurations for vulnerable services and modules; restrict and monitor management interfaces; watch logs for exploitation.
Reference Links
- Atlassian Multiple Products RCE and Critical Vulnerabilities Including CVE-2026-4800 (CVSS 9.8)
Overview
- Atlassian addressed 318 vulnerabilities across Bamboo, Bitbucket, Confluence, Crowd, Fisheye/Crucible, and Jira (including Jira Service Management); highest severity CVSS 9.8.
- Issues include RCE, path traversal, broken authentication and session management, server-side template injection, and third-party library flaws in Spring, Tomcat, Lodash, and Jackson Databind. No active exploitation or public exploit confirmed.
Impact
- Critical Lodash RCE in Confluence Data Center (CVE-2026-4800) allows unauthenticated or low-privileged code execution.
- RCE via CRLF injection in Spring Server-Sent Events affects Bamboo and Crowd (CVE-2026-47890); path traversal and file inclusion in Spring WebMVC; Tomcat authentication/session flaws; weak certificates in Data Center Docker images.
- SSTI in Bamboo, OS command execution in Fisheye/Crucible, stored XSS in Jira, and broken access control in JSM. Unpatched instances risk exposure of source code, build systems, wikis, and authentication services.
Affected / Fixed Versions
- Bamboo: affected 12.1.0–12.1.12 (LTS), 12.0.x, 11.0.x, 10.2.0–10.2.24 (LTS) and earlier; fixed in 13.0.1, 12.1.13 (LTS), 10.2.26 (LTS).
- Bitbucket: fixed in 10.5.2, 10.2.9 (LTS), 9.4.27 (LTS). Confluence: affected 11.0.0, 10.2.0–10.2.19 (LTS), 9.2.x LTS, 8.5.x LTS; fixed in 11.0.1, 10.2.20 (LTS), 9.2.27/9.2.28 (LTS).
- Crowd: fixed in 7.2.6. Fisheye/Crucible: affected 4.9.0–4.9.15; fixed in 4.9.16. Jira/JSM: fixed in 11.3.13 (LTS), 10.3.27 (LTS).
- Official Atlassian Docker images require updating to fix weak certificates (CVE-2026-91403).
Recommendations
- Upgrade all affected Atlassian products to the latest fixed versions immediately, prioritizing critical and internet-facing systems.
- Update Docker container images to the latest official versions.
Reference Links
- Puppet Enterprise Deserialization RCE CVE-2026-115731 (CVSS 9.4, Actively Exploited)
Overview
- Critical RCE in Puppet Enterprise and Puppet Core from improper deserialization of untrusted data in Puppet Server’s report-processing feature.
- An attacker with a CA-signed agent certificate can submit a crafted report to execute code under the Puppet Server service account. Environments with CA autosigning lacking strong verification are at higher risk, as unauthenticated attackers may obtain valid certificates.
- Listed in the KEV Catalog, confirming active exploitation.
Impact
- Remote code execution on Puppet Server with service account privileges and potential full control of affected instances.
Affected / Fixed Versions
- Affected: Puppet Enterprise 2026.0.0 before 2026.0.1; 2025.0.0 before 2025.12.1; 2023.8.0 before 2023.8.13; Puppet Core 9.0.0 before 9.1.3; Core 8.x before 8.22.2.
- Fixed: Puppet Enterprise 2026.0.1 (Core 9.1.3); 2025.12.1 and 2023.8.13 (Core 8.22.2); Puppet Core 9.1.3 and 8.22.2.
Recommendations
- Immediately upgrade all affected installations; prioritize internet-facing and critical Puppet Servers.
- Restrict CA autosigning to enforce strong verification of agent certificate requests.
- Audit Puppet Server logs and certificate activity for compromise.
Reference Links
- ARTEX Autonomous Penetration Testing Framework Weaponization and Supply-Chain Compromise Campaigns
Overview
- ARTEX is an open-source, LLM-driven multi-agent autonomous penetration testing framework written in Go, weaponized since at least August 2026 to automate intrusions from reconnaissance through exfiltration with minimal human involvement.
- It often reaches privileged access in under three hours and is used across financial, government, technology, defense, semiconductor, energy, healthcare, manufacturing, telecom, and legal sectors.
- Notable attacks: exfiltration of 67,000 individuals’ data from South Korean financial institutions and supply-chain compromises affecting over 2,250 ConnectWise Cloud Backup tenants.
- The original GitHub repository was removed in October 2026, but forks, mirrors, and a commercial variant (ARTEX PRO) remain accessible. Over 1,600 exposed ARTEX console instances were identified as of early November 2026.
Impact
- Accelerated, large-scale multi-target intrusions exploiting overlooked peripheral surfaces: partner/supplier portals, mobile backends, employee apps, legacy services, and third-party integrations.
- Supply-chain exploitation enabling lateral movement from weaker vendors into enterprise networks; compressed detection and response windows; significant exfiltration and credential abuse.
Technical Details and Tactics
- ARTEX uses external LLMs (Anthropic Claude, OpenAI GPT, DeepSeek, GLM, Grok, Kimi) to translate natural-language goals into automated attack steps run by specialized agents.
- Reconnaissance includes automated Azure permission mapping (SAS tokens, Service Principals, Kudu, Storage Account keys) with DryRun permission testing, web directory fuzzing (>500 requests), WHOIS lookups, active vulnerability scanning, and geolocation queries via api.db-ip.com.
- MITRE ATT&CK: T1595 (Active Scanning), T1580 (Collect Application Access Tokens), T1087.004 (Cloud Account Discovery), T1069.003 (Permission Groups Discovery), T1526 (Cloud Service Discovery).
Indicators of Compromise (IOCs)
- ARTEX hosts, proxy IPs, and C2 servers, including IPs linked to the South Korean financial campaigns and ConnectWise supply-chain compromise.
- Domain xcai.pro used as an LLM API reseller in campaigns.
- IPs include 38.244.50[.]120, 144.172.64[.]60-199-128, 187.15.88[.]186-217-89, and 199.19.95[.]66-68, among others in public telemetry.
Recommendations
- Inventory and secure internet-exposed applications, focusing on peripheral systems and vendor integrations.
- Restrict access to cloud admin interfaces, SCM/Kudu endpoints, hypervisors, and backups to trusted networks or private connectivity; disable unnecessary publishing authentication; use run-from-package deployment.
- Adopt Zero Trust access; enforce least privilege on multi-tenant app permissions; monitor app-only and service-principal activity in Microsoft 365.
- Review Entra, Purview, and SIEM audit logs for unauthorized credential or permission changes; revoke or rotate suspicious items.
- Apply rate limits, abuse controls, and anomaly detection on business-critical APIs; enforce strict server-side authorization on partner and employee portals; avoid exposing error details to unauthenticated users.
- Scan repositories and pipelines for embedded secrets and rotate them immediately.
- Preserve forensic evidence (sign-in logs, audit records, Graph telemetry) before containment; investigate ConnectWise anomalies back to October 2026; keep affected ConnectWise app connections disabled until restoration guidance and avoid deleting apps prematurely.
Reference Links
- Juniper Junos OS Evolved Linux Kernel IPv4 TCP Use-After-Free CVE-2026-98432 (CVSS 9.8)
Overview
- Critical use-after-free in the Linux kernel’s IPv4 TCP connection handling affects all versions of Junos OS Evolved. An unauthenticated remote attacker can send crafted TCP SYN packets with IPv4 options to reach freed kernel memory.
Impact
- Kernel memory disclosure, denial of service, and potential remote code execution.
Affected / Fixed Versions
- Affects all Junos OS Evolved versions. No fixed releases yet.
- Workaround: a Python 3 script installing an iptables rule to drop TCP SYN packets with IPv4 options on non-iri interfaces.
Recommendations
- Deploy the iptables workaround on all affected devices per Juniper advisory JSA119568.
- Monitor for suspicious SYN packets with IPv4 options; confirm the workaround persists after reboots and RE switchovers.
- Upgrade once fixed releases are available.
Reference Links
- Oracle Fusion Middleware and WebLogic Server Multiple Critical Vulnerabilities Including CVE-2026-114938 (Java4Shell) with Public PoC
Overview
- Oracle’s November 2026 Critical Security Patch Update addresses 655 vulnerabilities across Fusion Middleware, WebLogic Server, Access Manager, PeopleSoft, E-Business Suite, and more.
- CVE-2026-114938 (Java4Shell) is a high-severity memory corruption flaw in the OpenJDK Base64 decoder enabling unauthenticated RCE, with a public PoC.
- Other critical flaws include unauthenticated remote compromise in Oracle Access Manager (CVE-2026-114835, CVE-2026-91500, CVE-2026-91510), WebLogic Server (CVE-2026-91569), and PeopleSoft (CVE-2026-111497, CVE-2026-114934).
Impact
- Potential full system compromise via unauthenticated RCE; confidentiality, integrity, and availability at risk.
- Broad impact across Oracle Database Server, Java SE, GraalVM, E-Business Suite, MySQL, Communications products, Enterprise Manager, PeopleSoft, Siebel, and Primavera.
Affected / Fixed Versions
- Detailed affected products and fixed versions are in Oracle’s advisory.
Recommendations
- Apply the November 2026 CSPU immediately; prioritize Java4Shell (CVE-2026-114938) given the public PoC.
- Monitor Oracle advisories for updates.
Reference Links
- Mitel MiCollab Audio, Web and Video Conferencing Command Injection MTLVULN-1864 (CVSS 10.0)
Overview
- Mitel addressed five vulnerabilities in MiCollab’s AWV component: command injection, XML injection, arbitrary file upload, unauthorized access, and authentication bypass.
- MTLVULN-1864 is a blind command injection allowing unauthenticated remote arbitrary command execution; MTLVULN-1862 permits XML injection leading to command execution. Chained exploitation could amplify impact.
Impact
- Unauthenticated command execution and potential full compromise of integrity, confidentiality, and availability; limited unauthorized information access and unauthorized actions via authentication bypass.
Affected / Fixed Versions
- Affected: MiCollab 10.3 (10.3.0.18) and earlier, AWV component. Fixed: patch via KB000129195; later releases include fixes.
Recommendations
- Apply the security patch or upgrade to a later fixed release.
- Temporarily restrict external access to the AWV component; follow KB000129195.
Reference Links
- GitLab CE/EE Multiple Authorization and XSS Vulnerabilities CVE-2026-105739 (Actively Exploited)
Overview
- Updates fix four high-severity flaws: XSS (CVE-2026-105739, CVE-2026-108153) and incorrect authorization in CI/CD pipeline jobs and workspaces (CVE-2026-18241, CVE-2026-97187).
- Medium and low issues affect the GitLab Agent Server, Maven Virtual Registry, Pages, Confluence integration, Merge Request Discussions, Package Registry, Duo Agent, and others, enabling JavaScript execution, unauthorized data access, privilege escalation, CI/CD manipulation, credential exposure, internal network access, and DoS.
Impact
- Authorization bypass, XSS, privilege escalation, and service disruption.
Affected / Fixed Versions
- Vulnerable: GitLab CE/EE before 19.5.6, 19.4.8, and 19.3.9. Fixed in those releases.
Recommendations
- Update immediately to 19.5.6, 19.4.8, or 19.3.9 as applicable.
- Review CI/CD pipeline and workspace access controls; monitor for JavaScript injection or unauthorized access.
Reference Links
- N-able N-central Multiple Vulnerabilities Including Local Privilege Escalation and RCE CVE-2026-129271
Overview
- Multiple flaws in N-central RMM/UEM: local privilege escalation, authentication bypass, syslog injection enabling root command execution, missing authentication/authorization, XXE, SSRF, path traversal leading to RCE, and audit log tampering.
- CVE-2026-129271: local privilege escalation in Patch Management. CVE-2026-129286: path traversal in SIS GenericFiles resulting in RCE.
Impact
- Authentication and authorization bypass, privilege escalation, root command execution, compromise of N-central servers and managed endpoints, unauthorized server-side requests, policy disclosure, user enumeration, and audit log manipulation.
Affected / Fixed Versions
- Affected: N-central before 2026.4 Hotfix 3. Fixed: 2026.4.1.39 (Hotfix 3) or later; Agents and Probes at 2026.4.1 HF3 or later.
Recommendations
- Upgrade servers, Agents, and Probes to 2026.4.1.39 (HF3) or later.
- Restrict administrative access, strengthen authentication and access controls, monitor for compromise, and follow N-able updates.
Reference Links
- https://status.n-able.com/2026/11/16/n-central-2026-4-hf3-release-notes/
- https://uptime.n-able.com/event/202669/
- Atlassian Confluence Data Center RCE CVE-2026-90436 (Actively Exploited)
Overview
- Critical unauthenticated RCE in Confluence Data Center from an unsafe reflection flaw in the bundled Apache Axis 1.x SOAP library, executing code with Confluence OS account privileges without authentication or interaction.
- Actively exploited in the wild, confirmed by Rapid7 on November 13, 2026.
Impact
- Unauthorized access, data modification, information disclosure, and service downtime.
Affected / Fixed Versions
- Affected: Confluence Data Center 1.x through 9.x when the Remote API (XML-RPC & SOAP) is enabled. Fixed: 9.2.28 and later; 10.x and beyond not affected.
Recommendations
- Upgrade to 9.2.28 or a later unaffected version; if not feasible, disable the Remote API (XML-RPC & SOAP).
- Inventory deployments, verify Remote API status, prioritize internet-facing instances, and investigate logs for exploitation.
- Review other internal apps using Apache Axis 1.x; migrate to Axis 2 or patch; monitor Atlassian and Apache advisories.
Reference Links
- NginxPulse Unauthenticated SQL Injection CVE-2026-129171 (CVSS 9.3)
Overview
- Critical unauthenticated SQL injection in NginxPulse via the /api/stats/:type endpoint. Exploitation can give unauthorized database access and file read/write on the server using PostgreSQL superuser functions.
Impact
- Remote code execution and data compromise; potential full server compromise via database superuser capabilities.
Affected / Fixed Versions
- Affected: NginxPulse 1.5.0 through 1.6.25. Fixed: 1.6.26 and later.
Recommendations
- Upgrade to 1.6.26 or later; inventory all deployments, including Docker and Docker Compose.
- Restrict access to management interfaces and API endpoints to trusted networks.
Reference Links
- Linux Kernel TCP Race Condition and Use-After-Free CVE-2026-99288 (NoisyTCP)
Overview
- Critical race condition in the kernel’s TCP implementation involving request socket reuse and concurrent packet processing, causing use-after-free and kernel memory corruption.
- Introduced October 2015, affecting Linux 4.4 and later; exploitable remotely without authentication via crafted TCP traffic. Affects many distributions and TCP applications such as Nginx, Apache, OpenSSH, and proxies. Upstream patch applied November 3, 2026 (commit 4e2b3ef19e0e).
Impact
- Remote kernel panic (DoS), potential kernel memory corruption, and disruption of containerized applications on the host.
Affected / Fixed Versions
- Potentially affected: Linux 4.4 and later, subject to vendor backports. Fixed stable releases: 6.12.117, 6.18.63, 7.3.4; upstream commit 4e2b3ef19e0e04adca38442f07dd6bc93c712bf1.
Recommendations
- Apply vendor kernel updates immediately, prioritizing internet-facing servers running TCP services.
- If patching is delayed: disable SYN cookies (sudo sysctl -w net.ipv4.tcp_syncookies=0 per network namespace; reduces SYN flood protection), and serialize to a single RX queue while disabling RPS and RFS on external interfaces.
Reference Links
- N-able N-central Multiple Critical Vulnerabilities Including XXE (CVE-2026-125883) and SSRF (CVE-2026-125881) with CVSS 9.3
Overview
- Critical and high-severity flaws in N-central: XXE, SSRF, OGNL injection, RCE, local privilege escalation, and stored XSS, affecting on-premises and hosted deployments.
Impact
- Unauthorized external resource access via XXE; unauthorized server-side requests via SSRF; RCE via OGNL injection and file repository upload path traversal; local privilege escalation via writable syslog-ng configuration and insecure IPC channels; persistent XSS via MMS3 patch metadata in administrator context.
Affected / Fixed Versions
- Affected: N-central before 2026.4 HF2 (build 2026.4.1.26). Fixed: 2026.4 HF2 (2026.4.1.26) and later.
Recommendations
- Upgrade to 2026.4 HF2 or later; update all Agents and Probes in hosted and on-premises environments.
- Disable the Legacy API; review logs, administrative actions, and file uploads for compromise; apply further hotfixes as released.
Reference Links
- HPE Networking Private 5G Core Multiple Vulnerabilities Including Prototype Pollution CVE-2026-33228 (CVSS 9.8)
Overview
- CVE-2026-33228: critical prototype pollution in NodeJS Flatted. CVE-2026-49759: high-severity stack-based buffer overflow in Erlang OTP crashing the BEAM VM. CVE-2026-95780: authentication and session-management weaknesses allowing access by previously authorized users.
- CVE-2026-95781 through CVE-2026-95784: high-severity unauthorized access, disclosure, or disruption. CVE-2026-95785: authenticated users can make unauthorized changes to system-managed content. CVE-2026-95786: medium-severity improper access control.
Impact
- Denial of service via crashes, unauthorized access and retained access, information disclosure, and integrity/availability impact via unauthorized content modification.
Affected / Fixed Versions
- Affected: HPE Networking Private 5G Core 1.26.1.4 and earlier. Fixed: 1.26.2.0 or later.
Recommendations
- Update to 1.26.2.0 or later.
Reference Links
- Zenitel IP Intercom Firmware and Connect Server OS Command Injection and Privilege Escalation CVE-2026-95377 (CVSS 9.4)
Overview
- Zenitel fixed 11 vulnerabilities in IP intercom firmware and Zenitel Connect, including unauthenticated OS command injection, privilege escalation, path traversal, and improper authorization.
- CVE-2026-95377: unauthenticated OS command injection via DHCP Option 66 in the EdgeAudioClient TFTP provisioning module. Zenitel Connect flaws can be chained to gain root on the server. No active exploitation reported.
Impact
- Same-segment attackers can execute root commands on intercoms; low-privileged authenticated attackers may run commands in application containers or escalate to host root; admins can write arbitrary files as root; improper JWT validation could enable PostgreSQL superuser roles, exposing encryption keys and password hashes.
Affected / Fixed Versions
- IP Intercom models VSFF6, VSFF8, VSFF8MP, VSFF8MN, VSFT with firmware before 9.6.3.0 (9.x) and before 10.1.3.0 (10.x); Zenitel Connect before 1.11.0.
- Fixed: 9.6.3.0+, 10.1.3.0+, and Connect 1.11.0+.
Recommendations
- Upgrade intercom firmware to 9.6.3.0/10.1.3.0 or later and Zenitel Connect to 1.11.0 or later.
Reference Links
- PostgreSQL Type Confusion and Multiple High-Severity Vulnerabilities Including CVE-2026-6460 (CVSS 8.8)
Overview
- PostgreSQL released fixes for 52 vulnerabilities plus over 140 bug fixes. CVE-2026-6460, a roughly 22-year-old type confusion flaw in composite data type handling, lets an authenticated user who can create or modify database objects execute arbitrary code with the PostgreSQL OS account’s privileges. Exploitation was demonstrated against PostgreSQL 18.6.
- Other high-severity flaws include unauthenticated DoS, use-after-free, SQL injection, privilege escalation, and arbitrary SQL execution.
Impact
- Full compromise under the PostgreSQL service account; DoS, privilege escalation, arbitrary SQL execution, and code execution via client utilities.
Affected / Fixed Versions
- Affected: 18.x before 18.7, 17.x before 17.12, 16.x before 16.16, 15.x before 15.20, 14.x before 14.25. Fixed: 18.7, 17.12, 16.16, 15.20, 14.25 (final release, end-of-life).
Recommendations
- Upgrade to the fixed minor releases immediately; rebuild or update all C-language extensions after the server upgrade.
- Update client utilities (pg_dump, pg_restore, psql, reindexdb, vacuumdb); reindex affected indexes; take a new full backup for incremental backup setups.
- Audit and restrict privileges to create or modify objects; verify third-party and custom extensions; monitor logs for suspicious schema changes.
- Plan migration off PostgreSQL 14 (end-of-life); confirm remediation with managed database providers.
Reference Links
- Check Point Gaia OS Local Privilege Escalation CVE-2026-123039 (CVSS 7.8, Actively Exploited)
Overview
- Local privilege escalation in the Gaia DynamicCLI process (dynamic_server) from an insecure world-writable IPC socket (/tmp/dynamic_socket); the root-privileged service does not authenticate local connections and executes received commands as root.
- An authenticated local user with code execution on an appliance can escalate to root. Confirmed active exploitation affecting about two dozen customers.
Impact
- Root command execution by local attackers and potential full compromise of Check Point appliances and management platforms.
Affected / Fixed Versions
- Affected: R82.20, R82.10, R82, R81.20, R81.10 (end of support), R81 (end of support).
- Fixed (Jumbo Hotfix Accumulators): R82.20 Take 14+; R82.10 Take 52+; R82 Take 132+; R81.20 Take 174+.
Recommendations
- Upgrade Security Gateways, Management and Log Servers, and Maestro Orchestrator to the fixed Jumbo Hotfix Accumulators.
- Review /var/log/dynamicCli.log for commands without the standard Clish wrapper, use of the system() sink, and strings such as “id;”, “whoami”, “/etc/shadow”, “ncat”, “nc -“, “/dev/tcp”, “curl “, “wget”, “base64”, “useradd”, “authorized_keys”.
Reference Links
- USN-9182-1: nginx Vulnerabilities
Overview
- Flaws in handling of certain configurations and protocol messages: try_files and alias directives, ClientHello parsing, and certain TLS connections.
Impact
- CVE-2026-125283: remote DoS (crash) from misconfiguration handling. CVE-2026-123640: remote information leak from ClientHello parsing. CVE-2026-118899: remote DoS from TLS connection handling.
Affected / Fixed Versions
- Fixes for Ubuntu 18.04, 20.04, 22.04, and 24.04 LTS where applicable.
Recommendations
- Update nginx to the fixed versions for your Ubuntu LTS release.
Reference Links
- USN-9181-1: OpenJDK 8 Vulnerabilities
Overview
- Improper authentication or authorization issues in OpenJDK 8 components (JGSS, JSSE, Hotspot, Core Libraries, Compiler).
Impact
- Remote attackers may read or modify sensitive data or cause denial of service.
Affected / Fixed Versions
- OpenJDK 8 packages updated with security and bug fixes.
Recommendations
- Update OpenJDK 8 packages; review advisories and changelogs for related changes.
Reference Links
- Hackers Exploit VMware vCenter Zero-Day Since September to Hijack Admin Accounts and Steal VM Data (CVE-2026-123890)
Overview
- China-linked ARMORY CASTLE has exploited a vCenter Server zero-day, CVE-2026-123890, since early September 2026, bypassing authentication to gain administrator access without the password. Initial network access came via a BeyondTrust Remote Support zero-day (CVE-2026-102764).
- Attackers copied vCenter directory data, created admin accounts, forged SAML tokens, stole encryption material and database credentials, and accessed files from Windows VM virtual disks.
Impact
- Compromise of central VMware management exposing multiple hosts and workloads; theft of Active Directory, registry, and DNS files; host-level access may evade endpoint detection on VMs; stolen signing keys enable admin impersonation and MFA bypass.
Affected / Fixed Versions
- Fixes: vCenter 9.1.1.0200, 9.0.2.0200, and 8.0 Update 3m; VMware Cloud Foundation 5.2.5 updates. No public fix for vCenter 7.0; contact support.
Recommendations
- Apply patches promptly; restrict management network access.
- Review logs for unusual admin activity, SSH access, new root SSH keys, and virtual disk access.
- Revoke sessions and replace compromised signing keys and credentials; extend incident response to affected Windows domains given stolen AD data.
Reference Links
- Sean Cairncross Out as National Cyber Director
Overview
- Cairncross is leaving the role coordinating federal cybersecurity policy; DoD CIO Kirsten Davies will serve as acting director.
- His office shaped AI security policy and critical infrastructure protection, influenced a June executive order encouraging voluntary government testing of leading AI models, and launched programs for water utilities in Texas and rural hospitals in South Carolina.
Impact
- Leadership transition amid focus on AI security, infrastructure protection, and ransomware against healthcare, signaling evolving federal priorities.
Recommendations
- Continue government-industry collaboration on AI security and resilience in underserved sectors.
- Fill key federal cybersecurity positions to ensure continuity.
Reference Links
- Windows 11 Narrator Lock Screen Zero-Day (FastUnlock) Bypasses Lock Screen Protections
Overview
- FastUnlock lets an attacker with physical access bypass parts of the lock screen and run code without signing in on a running PC, even with BitLocker enabled; it does not break BitLocker encryption or decrypt a powered-off drive.
- Tested on Windows 11 26H2 with November 2026 updates (Windows 10 not tested). It abuses the account recovery screen, where Narrator reaches controls outside the lock screen. No Microsoft advisory, CVE, or patch yet.
Impact
- Code execution under the non-administrator WsiAccount at the lock screen; no takeover of signed-in sessions or escalation to admin. Sleeping devices remain vulnerable since BitLocker startup PINs are not required on wake.
Recommendations
- Prefer full shutdown or properly configured hibernation over sleep; use a TPM plus startup PIN for BitLocker on at-risk devices.
- Limit physical access, treat locked but powered-on laptops with caution, and watch for Microsoft advisories.
Reference Links
- VMware Patches vCenter Zero-Day Exploited in Attacks Since September
Overview
- Broadcom released updates for a critical authentication bypass in vCenter Server, actively exploited by Chinese-linked hackers since September 2026.
Impact
- Attackers can bypass authentication and potentially gain unauthorized access to vCenter environments.
Recommendations
- Apply Broadcom’s latest security updates.
Reference Links
- ASUS Warns of Critical Hard-Coded Key Flaw in Control Center Express
Overview
- ASUS released Control Center Express v1.7.29 to fix three critical vulnerabilities in the management console, allowing remote attackers to potentially gain administrative control of the console and managed devices.
Impact
- Remote takeover of the management console and managed devices.
Affected / Fixed Versions
- Fixed: Control Center Express v1.7.29.
Recommendations
- Update to v1.7.29 or later.
Reference Links
- phpAfterShell Bypasses PHP2Shell Fix to Enable Remote Code Execution
Overview
- phpAfterShell, disclosed by Depthfirst, enables unauthenticated RCE on PHP servers and bypasses the October PHP2Shell patches. It involves PHP’s default file-based session handler and the upload progress feature: interrupted session file writes let PHP accept corrupted session files, leading to object deserialization and memory corruption.
- Applications using Redis or database session storage are not affected.
Impact
- RCE with PHP worker permissions, potentially exposing credentials, application files, or customer data.
Affected / Fixed Versions
- Fixed in PHP 8.2.36, 8.3.37, 8.4.28, and 8.5.13. October PHP2Shell fixes do not mitigate it. PHP 8.0 and 8.1 are out of support.
Recommendations
- Upgrade PHP to the fixed versions; if delayed, set session.upload_progress.enabled = 0 and restart PHP workers.
- Verify the PHP version serving web requests (not just CLI); rebuild and redeploy containers; check distribution advisories.
Reference Links
- PHP Security Updates Fix 18 Flaws, Including Critical RCE Bugs
Overview
- The PHP team released updates for 18 vulnerabilities; several are critical and let unauthenticated attackers remotely execute code.
Impact
- Unauthenticated remote code execution.
Recommendations
- Apply the latest PHP security updates promptly.
Reference Links
- Identity Context Is Essential for Exposure Management: Tenable One Adds Ways to Bring in Identity Security Data
Overview
- Tenable One now ingests third-party identity findings (stale accounts, misconfigurations, excessive privileges) via a dedicated Semperis Lightning connector and a customizable open connector.
- Identity security is often siloed, so attackers can chain identity flaws with other vulnerabilities. Semperis Lightning monitors Active Directory and Entra ID and feeds severity-rated findings into Tenable One.
Impact
- Identity risk is prioritized alongside other exposures; consolidated findings reduce the chance of ignoring identity exposures; continuous synchronization of findings and exposure state.
Recommendations
- Incorporate identity security data into exposure management programs and use Tenable One connectors.
- Prioritize high-severity identity exposures alongside other vulnerabilities.
Reference Links
- USN-9175-1: Linux Kernel (Oracle) Vulnerabilities
Overview
- Security issues across drivers, architectures (ARM64, PowerPC, RISC-V, x86), MediaTek and Mellanox network drivers, USB, Ext4 and BTRFS, SELinux and AppArmor, and other kernel components; fixes cover CVEs from 2024 to 2026.
Impact
- Exploitation could lead to system compromise.
Recommendations
- Apply the Ubuntu Linux kernel update promptly.
Reference Links
- PHP Fixes 7 Critical Flaws, Including Bugs That Could Allow Remote Code Execution
Overview
- The PHP Foundation fixed seven critical vulnerabilities that could enable RCE, covering all supported PHP versions under certain conditions.
Impact
- Remote arbitrary code execution and potential system compromise.
Recommendations
- Apply the released security updates promptly.
Reference Links
- Drupal Patches Critical Installer Flaw That Could Give Unauthenticated Attackers Control of Sites
Overview
- A critical Drupal core flaw lets unauthenticated attackers use the built-in installer to attempt a site reinstall, leading to takeover or RCE. It affects uncommon configurations; fixes are available for all supported branches.
Impact
- Potential site takeover and RCE.
Recommendations
- Apply the fixed Drupal releases as soon as possible.
Reference Links
- ClamAV Vulnerabilities Affecting Cisco Products: November 2026
Overview
- Multiple ClamAV flaws let remote attackers bypass configured file and scan time limits, causing DoS that interrupts scanning. Cisco Secure Endpoint Connector on Windows, Linux, and Mac is affected; Secure Endpoint Private Cloud is not.
- Security Impact Rating is High on Windows (privileged context) and Medium on Linux and Mac.
Impact
- DoS affecting scanning operations; elevated risk on Windows due to privileged process execution.
Affected / Fixed Versions
- Cisco has released software updates.
Recommendations
- Apply Cisco’s updates to Secure Endpoint Connector on all affected platforms. No workarounds.
Reference Links
- Government Tech Debt Is a National Security Risk
Overview
- The FBI has disrupted multiple Chinese state-sponsored attacks on U.S. critical infrastructure (PlugX, Flax Typhoon, Volt Typhoon). They often exploit federal networks hampered by outdated, fragmented IT; agencies run multiple technology generations without clear integration or security mapping.
- About 80% of the federal IT budget goes to maintaining and securing existing systems. Adversaries use agentic AI to automate reconnaissance and speed exploitation.
Impact
- Fragmented, outdated infrastructure raises penetration risk and reduces resilience in crises; slow modernization and procurement hinder defense and recovery.
Recommendations
- Prioritize federal cyber investment and modernization with urgency comparable to defense acquisition; streamline procurement for trusted security technologies.
- Enforce vendor accountability in contracts and reduce platform duplication to shrink the attack surface.
Reference Links
- How CISA Is Prioritizing Companies for Talks About Security Improvements
Overview
- CISA identified about 100 critical infrastructure companies for system hardening discussions, based on importance to the economy, public health, safety, and security, with input from the Homeland Defense Working Group and intelligence on what adversaries value (pre-positioning and R&D targets).
- Engagements span interdependent entities such as port operators, energy, and telecom, focusing on resilience, minimum viability, and collective investment per the “CISA 2030” strategy, with CI-Fortify assessments emphasizing isolation and recovery.
Impact
- Coordinated prioritization is expected to improve critical infrastructure resilience.
Recommendations
- Encourage collaborative planning across interconnected sectors; use assessment insights to develop isolation and recovery strategies; keep refining investment and resilience priorities.
Reference Links
- Huntress Delivers 24/7 Response for Microsoft Defender Incidents Across Endpoints and Microsoft 365
Overview
- Huntress’s platform integrates Defender for Identity, Office 365, Cloud Apps, and Entra ID Protection data, revealing a token-replay campaign affecting 97 accounts across 79 organizations.
Impact
- Improved visibility and detection of multi-organization token-replay attacks.
Recommendations
- Consider integrated response platforms that consolidate endpoint and cloud identity alerts.
Reference Links
- Cisco Informational for Emergency Distrust of eMudhra Root and Intermediate Certificate Authorities
Overview
- In October 2026 a researcher showed that private keys for eight eMudhra intermediate certificates were weak due to a flawed April 2025 Key Generation Ceremony, allowing brute-forcing within seconds. Cisco found no affected products beyond those listed.
Impact
- Potential compromise of certificates from the vulnerable CAs, undermining trust and enabling man-in-the-middle attacks or impersonation.
Recommendations
- Follow the advisory for updates and mitigation; monitor Cisco and eMudhra guidance; replace affected certificates and distrusted CAs as outlined.
Reference Links
- CVE-2026-84869 in ScreenConnect Exploited to Run Malicious Files on MSP Technician Accounts
Overview
- Huntress confirmed active exploitation of CVE-2026-84869 against MSP technicians via ScreenConnect; the attack chain and IOCs are detailed.
Impact
- Attackers can run malicious files on MSP technician accounts, potentially compromising MSP environments and their clients.
Recommendations
- Review the IOCs and apply recommended mitigations.
Reference Links
- From View-Only Access to Container Root: Five Vulnerabilities in ONLYOFFICE Docs (Fixed)
Overview
- Rapid7 reported five flaws letting an attacker with view-only document access chain exploits to root OS command execution in the official DocumentServer container: native file read and Session JWT forgery, conversion parameter injection leading to code execution, path traversal with arbitrary file write, root-helper command injection, and view-only export of non-rendered document data.
- Tested on container 9.4.0; fixes confirmed in Document Server 10.0. A PoC is on GitHub; findings were responsibly disclosed.
Impact
- Escalation from view-only access to root inside the container; stealing the JWT signing key enables forging credentials; unauthorized file access; exfiltration of hidden content despite disabled downloads.
Affected / Fixed Versions
- Tested: onlyoffice/documentserver:9.4.0. Fixed: Document Server 10.0.
Recommendations
- Upgrade to Document Server 10.0 or later; review JWT key management and validation.
- Restrict access to /docbuilder, ensure proper permission checks, and monitor logs for suspicious access.
Reference Links
AI SOC
- What Is Preemptive MDR? A Guide for Security Leaders
Overview
- Preemptive MDR combines exposure context, threat intelligence, AI-assisted investigation, and human expertise for proactive, prioritized operations, connecting asset and vulnerability data to investigations to judge urgency and guide remediation.
- AI speeds investigation by retrieving, correlating, and organizing logs and activity in parallel, agentic-style analysis; human analysts remain central for validation and oversight.
Impact
- Better prioritization of weaknesses and hunting; clearer ownership and remediation tracking; faster containment and mitigation through AI-assisted, expert-validated workflows.
Recommendations
- Require provider transparency on how exposure context shapes investigations and how AI is used with oversight.
- Define incident response and remediation ownership; regularly review evidence of exposures resolved before exploitation.
Reference Links
- Why AI-ROC Is the Future of Exposure Management
Overview
- Vulnerability disclosures are surging, accelerated by frontier AI models, and attackers exploit within hours; manual handoffs and ticketing make remediation too slow.
- Qualys’s AI-native Risk Operations Center (AI-ROC) uses three agents: Agent Insta (near real-time detection without rescanning), Agent Val (production-safe exploit validation), and Agent Sara (automated, low-risk remediation with rollback and patch reliability scoring).
Impact
- Compresses weeks-long SLAs to hours or days; narrows millions of findings to confirmed exploitable ones; greater remediation confidence; a unified platform without tool handoffs.
Recommendations
- Adopt AI-ROC-style models integrating agents across detection, validation, and remediation.
- Use real-time detection, exploit validation, and automated remediation with patch reliability scoring, rollback, and policy-controlled deployment waves.
Reference Links
- Fix Tenable One Findings Where Developers Work: Tenable and OpenAI Daybreak Defense Factory
Overview
- Tenable One findings can be imported into OpenAI Daybreak Defense Factory via a local Codex Security CLI, preserving description, solution, severity, state, file paths, and line numbers.
- The open-source OpenAI connector uses the Inventory Export API or CSV exports; Daybreak helps developers prioritize, investigate, and propose fixes.
Impact
- Less context loss and duplicated triage between security and development; faster, more accurate remediation of code issues.
Recommendations
- Use the Daybreak Defense Factory connector and OpenAI’s setup guide; refresh imports regularly.
Reference Links
- Security for the Agentic Era, Powered by Microsoft Perception
Overview
- AI is compressing exploit development from months to minutes, and attackers use autonomous agents to run whole campaigns.
- Microsoft’s cyber stack orchestrates red, blue, and green team agents through Microsoft Perception for attack simulation, automated triage and investigation, and closed-loop remediation, alongside the Integrated SOC in Defender (SIEM, UEBA, threat intelligence).
- New initiatives provide playbooks, vulnerability shielding, and prioritized exposure reduction across multiple clouds.
Impact
- Higher urgency for rapid detection and response; automated, collaborative SOC functions augment analysts; addresses risk from AI-powered attackers and over-privileged agents that can be compromised.
Recommendations
- Adopt agentic security operations with coordinated agents and automated playbooks.
- Use AI Readiness Initiatives and vulnerability shielding; leverage Defender, Sentinel, and the wider ecosystem; keep human oversight over high-impact automated actions.
Reference Links
- Tenable One Adversary View Brings Claude Mythos’ Adversarial Reasoning to Your Environment
Overview
- Adversary View, via Tenable Hexa AI, uses Claude Mythos for adversarial reasoning over existing scan data, building an on-demand asset graph, identifying vulnerability chains and choke points, and prioritizing fixes that disrupt the most attack paths.
- It offers remediation guidance (patches, configuration changes), Jira and ITSM integration, scoping by tags, IP ranges, or business units, and interactive follow-up.
Impact
- Prioritizes by attack path disruption rather than severity alone; focuses on highest-leverage fixes; demonstrated multi-hop paths such as EternalBlue scenarios.
Recommendations
- Use Adversary View for on-demand attack path analysis, scoped to critical assets or business units; integrate ticketing to track remediation.
Reference Links
AI Threat Landscape
- Navigating Vulnerability Hysteria in the AI Era
Overview
- A Reporters’ Notebook episode discusses the aftermath of recent high-severity vulnerability disclosures, how fear of AI-assisted exploits drives panic and operational disruption, and growing warning fatigue among security professionals.
Impact
- Panic-driven responses create operational challenges; warning fatigue can reduce vulnerability management effectiveness.
Recommendations
- Calibrate responses to AI-related disclosures to avoid overreaction; improve alert prioritization and filtering.
Reference Links
- 30 Years of Gatekeeper Asymmetry Doom: The Paradox That Makes Powerful LLMs More Vulnerable
Overview
- Check Point Research describes “gatekeeper asymmetry doom”: less powerful vetting systems are consistently outmatched by the systems they evaluate, from hash-based antivirus to static binary analysis and pattern-matching tools like YARA.
- Even frontier LLMs face inherent limits vetting malicious code due to computational theory, and attackers exploit this with polymorphism and obfuscation. Recommends narrowing the power gap between gatekeepers and targets.
Impact
- Persistent difficulty defending against inputs more powerful or obfuscated than the vetting mechanism, exposing static detection and AI-powered analysis to sophisticated evasion.
Recommendations
- Recognize “doomed” attack surfaces and increase gatekeeper power or use it more effectively.
- Keep improving AI and vetting systems while understanding their theoretical and practical limits.
Reference Links